Inspiration
We have the ability to build everyone a JARVIS personal assistant to control things with your voice, and we believe your JARVIS personal assistant should be open-source and private.
What it does
Converts voice into text for Mac & Windows users. It is designed to work universally across all text box applications, including Slack, Notion, Google Docs, and Cursor.
In addition to simple dictation, the AI allows the user to verbally express intent, and crafts smart, polished text content in response. Users can activate Ito with a hotkey to provide hands-free operation.
Ito is built to increase productivity and simplify work by enabling users to think and speak, rather than type. It can be used to write emails, code, product briefs, Slack messages, meeting agendas, and social media posts.
Can draft full content from a single instruction given by the user. Ito is also customizable. Users can add custom vocabulary and configure hotkeys according to their preferences.
It has learning capabilities that enable it to improve and match a user's style based on previous conversations, contributing to a more personalized user experience.
How I built it
After vibe coding the first version, we tore it down and rebuilt it from scratch focused on reliability and performance.
Challenges I ran into
Lots of debugging with external keyboards and microphones, but overall it's led to more reliability.
Accomplishments that I'm proud of
I dictated this entire submission.
What I learned
This is the first native Mac OS and Windows app that we've developed. So there's a lot of learnings around building native for those platforms.
What's next for Ito
We're going to dive deeper into context engineering to increase accuracy based on things we can deduce from people's actions and behavior. All of this will be stored locally.
Log in or sign up for Devpost to join the conversation.